Update exporter force_distinct feature to consider the whole iterable

[bruth]

Currently, the implementation of the force_distinct flag in the exporter class will only remove consecutive duplicate rows. Depending on the sorting method, this could still produce duplicate rows later in the result set which is undesirable.

One caveat to this is that result set will be kept in memory for duration of the export. In practice, this is rarely an issue since the size of iterables are relatively small, but for large exports this could be very slow and potentially cause OutOfMemory errors.